Interactions between Hofmeister Anions and the Binding Pocket of a Protein

CitationFox, Jerome M., Kyungtae Kang, Woody Sherman, Annie Héroux, G. Madhavi Sastry, Mostafa Baghbanzadeh, Matthew R. Lockett, and George M. Whitesides. 2015. “Interactions Between Hofmeister Anions and the Binding Pocket of a Protein.” Journal of the American Chemical Society 137 (11) (March 25): 3859–3866. doi:10.1021/jacs.5b00187.

Published Versiondoi:10.1021/jacs.5b00187

Permanent linkhttp://nrs.harvard.edu/urn-3:HUL.InstRepos:25042504

Terms of UseThis article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Open Access Policy Articles, as set forth at http://nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#OAP

Share Your StoryThe Harvard community has made this article openly available.Please share how this access benefits you. Submit a story .

Accessibility

  1  

Interactions between Hofmeister Anions and the Binding Pocket of a Protein

Jerome M. Foxa, Kyungtae Kanga, Woody Shermand, Annie Hérouxe, Madhavi Sastryd, Mostafa

Baghbanzadeha, Matthew R. Locketta†, and George M. Whitesidesa,b,c*

a Department of Chemistry and Chemical Biology, Harvard University

12 Oxford Street, Cambridge, MA 02138, USA. b Wyss Institute for Biologically Inspired Engineering, Harvard University, Cambridge, MA

02138, USA. c The Kavli Institute for Bionano Science and Technology, Harvard University, Cambridge, MA

02138, USA. d Schrödinger, Inc., 120 West 45th Street, New York, NY 10036, USA.

e National Synchrotron Light Source, Photon Sciences Directorate, Brookhaven National Lab,

Building 745, Upton, NY 11937, USA.

  2  

ABSTRACT

This paper uses the binding pocket of human carbonic anhydrase II (HCAII, EC 4.2.1.1)

as a tool to examine the properties of Hofmeister anions that determine (i) where, and how

strongly, they associate with concavities on the surfaces of proteins and (ii) how, upon binding,

they alter the structure of water within those concavities. Results from X-ray crystallography and

isothermal titration calorimetry show that most anions associate with the binding pocket of

HCAII by forming inner-sphere ion pairs with the Zn2+ cofactor. In these ion pairs, the free

energy of anion-Zn2+ association is inversely proportional to the free energetic cost of anion

dehydration; this relationship is consistent with the mechanism of ion pair formation suggested

by the “Law of Matching Water Affinities.” Iodide and bromide anions also associate with a

hydrophobic declivity in the wall of the binding pocket. Molecular dynamics simulations suggest

that anions, upon associating with Zn2+, trigger rearrangements of water that extend up to 8 Å

away from their surfaces. These findings expand the range of interactions previously thought to

occur between ions and proteins by suggesting that (i) weakly hydrated anions can bind

complementarily shaped hydrophobic declivities, and that (ii) ion-induced rearrangements of

water within protein concavities can (in contrast with similar rearrangements in bulk water)

extend well beyond the first hydration shells of the ions that trigger them. This study paints a

picture of Hofmeister anions as a set of structurally varied ligands that differ in size, shape, and

affinity for water and, thus, in their ability to bind to—and to alter the hydration structure of—

polar, nonpolar, and topographically complex concavities on protein surfaces.

  3  

INTRODUCTION

The non-covalent association of simple ions and proteins in aqueous solution plays a

central role in many of the biochemical processes that constitute “life.” By binding and

transporting Na+, K+, Mg2+, Ca2+, SO42-, HCO3

-, and Cl-, ion channels in cell membranes regulate

intracellular volume and pH,1,2 control the uptake of nutrients and the release of metabolites,3–5

engage in signal transduction,6,7 and mediate action potentials;8,9 by associating with—and

subsequently oxidizing—I-, thyroid peroxidases enable the production of essential iodine-

containing hormones; 10 and by binding inorganic phosphate (and longer chain phosphate esters)

kinases and phosphatases regulate the activity of enzymes and receptors throughout the cell.11

Despite their importance in a range of biochemical phenomena, however, ion-protein interactions

in aqueous environments remain incompletely understood at the molecular level.12–17

Two questions summarize existing uncertainty concerning the mechanisms by which ions

and proteins interact in aqueous systems: (i) What attributes of ions and the surfaces of proteins

determine where, and how strongly, they associate with one another? (ii) How do ions alter the

structure of water solvating those surfaces (which differ in charge, topography, and organic

functionality)? Answering the first question would explain why proteins exhibit different

affinities for ions of the same charge (e.g., Na+ vs. K+).18–20 Answering the second question

would explain how ions, by reorganizing the water solvating protein substructures (e.g.,

declivities, charged elements, polar and nonpolar surfaces), alter the interactions in which those

substructures participate.21–24

This study addresses these two questions by examining ion-protein interactions in an

experimentally well-defined model system: the binding pocket of human carbonic anhydrase II

(HCAII, EC 4.2.1.1).25,26 Using isothermal titration calorimetry (ITC), X-ray crystallography,

  4  

and molecular dynamics simulations, we examined the association of Hofmeister anions with the

binding pocket of HCAII, and with the molecules of water filling that pocket. This binding

pocket is a good model system for studying non-covalent interactions between ions and proteins

for two reasons: (i) It has both a polar surface (Asn-62, His-64, Asn-67, Gln-92, Glu-206)27 and a

nonpolar surface (Phe-131, Val-135, Leu-198, Pro-201, Pro-202, Leu-204).28 (ii) It contains a

positively charged metal cofactor (Zn2+) that can associate with anions that occupy different

positions in the Hofmeister series (e.g., SO42-, CH3COO-, Cl-, Br-, NO3

-, I-, SCN-).29–33

The Hofmeister series ranks the influence of ions on a wide variety of physical processes,

most notably, their tendency to precipitate proteins from aqueous solution (Figure 1A, Appendix

1 of the SI).16,34 We reasoned that anions with different positions in this series might exhibit

different propensities to (i) partition into the binding pocket of HCAII (by interacting with the

Zn2+ cofactor and, perhaps, polar and nonpolar residues) and (ii) reorganize molecules of water

filling that pocket. By examining the association of Hofmeister anions with the binding pocket of

HCAII, we hoped to identify attributes of ions that influence (i) where, and how strongly, they

bind concavities on the surfaces of proteins and (ii) how, upon binding, they perturb the local

structure of water.

Background: Key Terms and Concepts. Figure 1A shows the Hofmeister series of

anions. Anions to the left of chloride, termed “kosmotropes”, tend to stabilize folded proteins

(relative to unfolded proteins), and cause proteins to precipitate from aqueous solution.16,34

Anions to the right of chloride, termed “chaotropes”, tend to promote denaturation, and enhance

the solubility of proteins in solution. Kosmotropes are generally small (e.g., radius < 1.8 Å for

monovalent anions)35 and strongly hydrated; chaotropes are generally large (e.g., radius > 1.8 Å

for monovalent anions) and weakly hydrated.18

  5  

We use the terms “strongly hydrated” or “weakly hydrated” to refer to the free energies

of hydration of various anions (ΔG°hydration, the free energy change associated with the transfer of

one mole of ion from the gas phase to water at standard state).36 For strongly hydrated anions,

values of ΔG°hydration are more negative (e.g., ΔG°hydration ≈ -90 kcal/mol for CH3COO-);36 for

weakly hydrated anions, values of ΔG°hydration are less negative (e.g., ΔG°hydration ≈ -50 kcal/mol

for ClO4-).

RESULTS AND DISCUSSION

Ion Pairs and the “Law of Matching Water Affinities.” Several studies have proposed

that ions associate with the surfaces of proteins by forming ion pairs in accordance with the

so-called “Law of Matching Water Affinities” (Appendix 2 of the SI).18,37–41 This qualitative

“Law” (or, perhaps, more appropriately, “empirically based hypothesis”) suggests that inner-

sphere ion pairs form preferentially between oppositely charged ions with similar free energies

of hydration. Two implications follow: (i) Small, strongly hydrated ions—ions for which ion-

water interactions are more free energetically favorable than water-water interactions—will

associate with one another because the free energetic cost of partially desolvating those ions is

more than compensated by the free energetic benefit of forming ion pairs. (ii) Large, weakly

hydrated ions—ions for which ion-water interactions are less free energetically favorable than

water-water interactions—will associate with one another because the free energetic cost of

partially desolvating those ions is more than compensated by the free energetic benefit of

forming additional water-water interactions.

Empirical support for the Law of Matching Water Affinities (as it pertains to ion-protein

interactions) is based, in part, on observations that ions and/or surface charges with similar levels

  6  

of hydration tend to associate with one another.14,39,41 For example, weakly hydrated anions (e.g.,

SCN-) tend to associate with the weakly hydrated side chains of lysine and arginine; strongly

hydrated anions (e.g., HPO42-) tend to associate with strongly hydrated cations (e.g. Ca2+) present

at low concentrations (10-7 M) within the cell.18,42 (Spectroscopic examination of the association

of divalent cations with carboxylate side chains of polypeptides indicate that this rule of thumb

might not hold for multivalent ions.43) The absence of corroborating thermodynamic

investigations, however, has left the mechanism of ion-pair formation implied by this theory (and

other theories) both (i) incompletely validated and (ii) without a predictive quantitative extension

(i.e., a simple rule, grounded in thermodynamics, capable of predicting the relative affinities of

two ions for a particular charged group).44 We tested the Law of Matching Water Affinities—and

evaluated a possible quantitative extension of this theory—by examining the correlation between

ΔG°hydration for Hofmeister anions and their affinity for a single charged element: the Zn2+

cofactor of HCAII.

Two States. We discuss the non-covalent association of anions and proteins by

comparing two states: an initial state, which consists of anions and proteins—not interacting with

each other—in aqueous solution, and a final state, which consists of anion-protein complexes in

aqueous solution (Figure 1B). Changes in thermodynamic properties resulting from anion-

protein association (ΔJ°bind, where J = G, H, or TS), thus, reflect a difference in thermodynamic

properties between the initial state and the final state (ΔJ°bind = Jfinal - Jinitial).

The Thermodynamic Basis of Association between Anions and the Zn2+ Cofactor.

Hofmeister anions bind Zn2+ too weakly (i.e., the free energy of binding is too small) to permit

the direct examination of anion-Zn2+ interactions with isothermal titration calorimetry (ITC). To

obtain values of ΔJ°bind (where J = G, H, or TS) for the association of anions and Zn2+ (Figure

  7  

1B), we thus employed a competition assay (Methods, Figure S1) similar to that employed by

Zhang et al. to study the binding of low-affinity ligands to the protein tyrosine phosphatase 1B

(EC 3.1.3.48).45 Briefly, using ITC, we measured the dissociation constant and enthalpy of

binding for the association of HCAII and benzo[d]thiazole-2-sulfonamide (BTA)—a high-

affinity ligand (Kd,BTA = 60 ± 30 nM) that binds the Zn2+ cofactor of HCAII46—in the presence

(and absence) of sodium salts of ten different Hofmeister anions (100 mM, Figure 1A). (We

note: in this discussion, values of Kd,BTA represent the pKa-corrected values corresponding to the

association of the deprotonated form of BTA with the water-bound form of HCAII. This

correction is detailed in the SI). In the presence of sodium salts, BTA displaces Zn2+-bound

anions, and the observed values of the dissociation constant (K!,!"#!"# ) and enthalpy of binding

(∆𝐻°  !"#$,!"#!"# )—that is, values estimated under the assumption that no ions are present—differ

from values of the dissociation constant (Kd,BTA) and enthalpy of binding (ΔH°bind-BTA) determined

in the absence of ions in accordance with Eqs. 1 and 2, where Kd,anion and ΔH°bind,anion are the

dissociation constant and enthalpy of binding, respectively, for a specific anion, and [Atot] is the

total concentration of that anion.

Kd ,BTAobs = Kd ,BTA +

Kd ,BTA

Kd ,anion

Atot[ ] (1)

ΔHbind ,BTAobs = ΔHbind ,BTA

! −ΔHbind ,anion

!

1+Kd ,anion

Atot[ ] (2)

For each anion, we used Eqs. 1 and 2 to determine Kd,anion and ΔH°bind,anion; from these values ,

we estimated ΔG°bind,anion and -TΔS°bind,anion (Figure 2A, SI).

Our results indicate that the chloride and the chaotropes engage in enthalpically favorable

(ΔH°bind,anion < 0), entropically unfavorable (-TΔS°bind,anion > 0) interactions with the the Zn2+

  8  

cofactor (Figure 2A). (We note: although anions may have additional binding sites in the region

of the binding pocket taken up by BTA, and associated binding events would be reflected in the

thermodynamic parameters measured with competition experiments, there is no crystallographic

evidence for such sites.21–24) Interestingly, from left to right across the Hofmeister series (i.e.,

with increasing chaotropicity of the anions), values of ΔH°bind,anion decrease, and values of -

TΔS°bind,anion increase with almost complete compensation, and values of ΔG°bind,anion decrease

only slightly (from -2.3 ± 0.1 kcal/mol for Cl- to -3.2 ± 0.1 kcal/mol for SCN-). This type of

enthalpy/entropy (H/S) compensation is believed to arise, in many bimolecular interactions, from

rearrangements in the molecules of water that solvate interacting species,47,48 and, thus, suggests

that anion-Zn2+ association is strongly influenced by thermodynamic contributions from

desolvation of the anion and/or Zn2+ cofactor. (We note: with calorimetry—although less with

ITC than with experimental methods that rely on Van’t Hoff analysis—errors in measured values

of ΔH°bind translate to errors in estimates of ΔS°bind, and can cause H/S compensation to be

perceived where it does not occur.49 We used a number of precautions, and carried out statistical

checks, to reduce such errors; see SI Methods).

Kosmotropes, in contrast with chaotropes, bind weakly to the Zn2+ cofactor (Figure 2A)

or, in the case of SO42- and HPO3

2-, not at all (i.e., too weakly to be detected under the conditions

of our experiments). For HCO3- and CH3COO-, values of ΔH°bind,anion and -TΔS°bind,anion again

nearly compensate one another, but not in a manner consistent with the trend exhibited by

chaotropes. This inconsistency likely arises from different mechanisms of binding. HCO3- is a

substrate of HCAII; CH3COO- is a substrate analogue. Unlike values of ΔJ°bind,anion for

chaotropes, values of ΔJ°bind,anion for HCO3- and CH3COO- involve contributions from hydrogen

bonds between the bound anions and amino acids near the Zn2+ cofactor.29,50

  9  

To examine the relationship between the affinity of specific anions for the Zn2+ cofactor

and the free energetic cost of anion desolvation, we plotted ΔG°bind,anion for each anion

(chaotropes and komostropes) against literature values36 of their free energies of hydration

(ΔG°hydration; Figure 2B). Values of ΔG°bind,anion decrease linearly with ΔG°hydration, and indicate

that anions most capable of shedding their first hydration shells bind most tightly to Zn2+. This

linear relationship, which suggests that the affinity of anions for the Zn2+ cofactor correlates

inversely with their affinity for water, is consistent with the mechanism of ion pair formation

implied by the Law of Matching Water Affinities.38

Evidence of Hydrophobic Interactions between Anions and HCAII. Modeling studies

by several groups have suggested that large, poorly hydrated anions can associate with nonpolar

regions on the surfaces of proteins.19,51–54 Experimental studies have substantiated these

predictions by demonstrating that weakly hydrated anions can associate with nonpolar

concavities in synthetic host systems;55,56 hydrophobic interactions between anions and the

surfaces of proteins, however, have proven difficult to examine experimentally, and the role of

hydrophobicity in ion-protein association in aqueous environments remains controversial.37,57

We used X-ray crystallography to search for hydrophobic binding sites for ClO4-, SCN-, I-, and

Br- in the binding pocket of HCAII. These anions are four of the most poorly hydrated included

in the present study (i.e., they have smaller values of ΔG°hydration than the other anions examined,

Table S6); thiocyanate, iodide, and bromide have the added advantage that they exhibit

anomalous scattering (due to S, I, and Br atoms)—an attribute that makes them useful tools for

the detection of secondary, low-occupancy binding sites.58,59

Structures of HCAII complexed with ClO4- and SCN- reveal a single ion in the binding

pocket—bound, in each case, to the Zn2+ cofactor. Both anions displace H2O-338, shift the

  10  

position of H2O-263, and leave Zn2+ in a pentacoordinated geometry (Figures 3A, and S3A-

S3B). By contrast, the structures of HCAII complexed with iodide and bromide show four

binding sites (Figures 3B-3D and S4D; Appendix 3 of the SI). Here, for simplicity of discussion,

we discuss the binding sites of iodide, which are identical to those of bromide, by referring to

them in order of their proximity to the Zn2+ cofactor (I-1 through I-4, closest to farthest away). I-

1 and I-2 denote alternative binding sites for ion-Zn2+ complexation and are not occupied

simultaneously; these likely permit the formation of an inner-sphere ion pair (one that involves

ion-ion contact) and an outer-sphere ion pair (one that involves a shared solvating water),

respectively. I-3 denotes a binding site at the border of the hydrophilic and hydrophobic surfaces;

it sits in close proximity (3.5 Å) to the amine of Gln-92 (Figures 3B and 3C). I-4 denotes a

binding site within a small hydrophobic declivity formed by five nonpolar side chains near the

mouth of the binding pocket (Figs. 3B and 3D). As there is no positive charge proximal to the I-4

site, and as analysis of the surface charge within this site (an analysis carried out with the

Adaptive Poisson-Boltzmann Solver60 package for PyMOL, Appendix 4 of the SI) shows little

excess positive charge, Coulombic attraction is not the primary driving force for the association

of iodide with this site.

The absence of secondary binding sites for the thiocyanate anion, which has a volume,

free energy of hydration, and polarizability nearly indistinguishable from those of the iodide

anion (Table S6),36 suggests that ion shape (a parameter rarely mentioned in discussions of ion-

protein association) may influence the ability of ions to engage in hydrophobic interactions. The

I-4 binding site, in particular, has a hemisphere-like shape that can easily accommodate spherical

iodide and bromide ions, but not a linear ion such as SCN- (Figures 3D and S5A-C).

  11  

The I-4 binding site provides direct evidence that poorly hydrated anions (i.e., iodide and

bromide) can associate with complementarily shaped hydrophobic declivities on the surfaces of

proteins. Previous molecular dynamics simulations provide evidence of an attraction between

chaotropic anions and nonpolar regions on protein-like polymers;51,52 here, crystallographic

evidence indicates that two chaotropes can associate directly with a binding site formed by five

nonpolar side chains. The existence of such a binding site suggests that theories of ion-protein

interactions focused exclusively on the formation of ion pairs may oversimplify the variety of

these interactions.

Anion-Induced Perturbations of the Structure of Water within the Binding Pocket.

Many studies have suggested that Hofmeister ions reduce or enhance the solubility of proteins—

a process termed “salting out” or “salting in,” respectively—by reducing or enhancing hydration

of solvent-exposed residues.23,24,61,62 The mechanisms and thermodynamic implications of such

adjustments to hydration, however, remain poorly understood. We examined ion-induced

perturbations of water structure inside the binding pocket of HCAII by using the WaterMap

method (Schrödinger Inc.,63–65 see SI Methods). WaterMap uses explicit-solvent molecular

dynamics simulations, and inhomogeneous solvation theory, to calculate the enthalpy, entropy,

and free energy of hydration sites within solvated proteins, relative to bulk water.66,67 Unlike X-

ray crystal structures, which reveal only the positions of well-ordered, highly localized (i.e.,

enthalpically stable) waters, WaterMap predicts the positions and thermodynamic properties of

all waters—well-ordered or otherwise—in a structure.

The association of anions with the Zn2+ cofactor of HCAII (the process depicted in Figure

1B) is coincident with rearrangements in the molecules of water filling the binding pocket. To

evaluate the thermodynamic contribution of these rearrangements to anion-Zn2+ association, we

  12  

summed the thermodynamic properties (enthalpy, entropy, and free energy) of hydration sites

located in the binding pockets of anion-bound (ΔJ°WM,HCA-anion) and native (ΔJ°WM,HCA) HCAII

complexes, and we calculated the difference of these sums (e.g., ΔJ°WM, anion = ΔJ°WM,HCA-anion

- ΔJ°WM,HCA, where WM denotes values calculated from WaterMap, and J = G, H, or TS; see SI

Methods). Crystal structures of HCAII containing a variety of Zn2+-bound anions (collected here

and elsewhere)21–24 allowed us to perform these calculations for anions spanning the Hofmeister

series (SI Methods). Results from our calculations suggest that anions, upon forming ion pairs

with Zn2+, bring about entropically favorable (-TΔS°WM,anion < 0) and enthalpically unfavorable

(ΔH°WM,anion > 0) rearrangements of water inside the binding pocket (Figure 4A). Interestingly,

values of ΔJ°WM,anion (where J = G, H, or TS) are similar across the Hofmeister series (Figure

4A). This result, in light of the linear relationship between the free energy of anion-Zn2+

association and the free energy of anion hydration (Fig. 2B), suggests that differences in the

affinity of Hofmeister anions for the Zn2+ cofactor are not the result of differences in anion-

induced rearrangements of water inside the binding pocket, but rather from differences in (i) the

free energetic cost of anion desolvation and (ii) the free energetic benefit of forming an anion-

Zn2+ pair.

The Length Scale of Anion-Induced Perturbations of Water within the Binding

Pocket. The results of several spectroscopy studies of ions in bulk water suggest that the effect

of ions on the structure of water is limited to their first hydration shells.68–70 Complementary

experimental examinations of ions adsorbed at interfaces, however, have remained difficult, and

the length scale over which ions perturb interfacial water remains unclear.71–73 Using results from

WaterMap calculations, we estimated the distance over which Zn2+-bound anions trigger

rearrangements of water within the binding pocket of HCAII by examining ΔH°WM,anion(d) and

  13  

-TΔS°WM,anion(d), the changes in enthalpy and entropy, respectively, that result from binding-

induced rearrangements of water that occur beyond a distance d Å from the surface of the Zn2+-

bound anion (Figure 4B). Figure 4C shows the representative case of thiocyanate (other ions

show similar trends; Figure S6); this figure indicates that rearrangements of water coincident

with the binding of SCN- persist well beyond the first hydration shell (~ 2.5 Å) of this anion, and

extend up to 8 Å away from its surface (beyond d = 8 Å, values of ΔJ°WM,anion(d) decrease to less

than 10 % of ΔJ°WM,anion). This distance suggests that the influence of anions on the structure of

water at protein/water interfaces—or, at least, within the declivities of proteins—can extend well

beyond the single hydration shells that demark the limit of their influence on the structure of bulk

water. This result is consistent with previous molecular dynamics simulations suggesting that

water within confined regions (e.g., the binding pockets of proteins) exhibits long-range

structure;63 alterations to the charge/structure of such regions are, thus, likely to have long-range

consequences (such as those depicted in Figure 4C).

The Influence of Rearrangements of Water on the Binding of Anions to the I-4 Site.

Hydrophobic interactions between ligands and proteins often involve the free energetically

favorable release of water from hydrophobic binding pockets.46,74,75 To examine the role of

displaced water in the association of iodide or bromide with the hydrophobic I-4 site, we used

WaterMap to estimate the thermodynamic properties of molecules of water filling the binding

pocket of HCAII in the presence and absence of bound iodide or bromide anions. Results suggest

that the binding of iodide and bromide (separately) is coincident with the displacement of two

molecules of water that are enthalpically and entropically unstable (relative to bulk water;

Figures 5 and S7); the association of iodide and bromide with the I-4 hydrophobic declivity,

thus, resembles the interaction of nonpolar ligands with hydrophobic binding pockets.46,75,76

  14  

CONCLUSION

Ion-Protein Association. This study uses the binding pocket of HCAII as a tool to

identify the properties of Hofmeister anions that determine (i) where, and how strongly, they

associate with concavities on the surfaces of proteins and (ii) how, upon binding, they alter the

structure of water within those concavities. We find that anions can associate with the binding

pocket of HCAII by forming inner-sphere ion pairs with the Zn2+ cofactor, and, in the case of

iodide and bromide, by associating directly with a hydrophobic declivity.

For anion-Zn2+ association, calorimetry and X-ray crystallography suggest that the free

energy of anion binding is inversely proportional to the free energetic cost of anion dehydration;

this relationship is consistent with the mechanism of ion pair formation suggested by the Law of

Matching Water Affinities and, thus, suggests that this theory may explain, in some biophysical

contexts, the relative affinity of anions for positive charges on the surfaces of proteins. The

formal extension of the Law of Matching Water Affinities to positive charges present in specific

contexts (e.g., charges within specific classes of concavities) will require calorimetric and

crystallographic studies of anion binding to pockets that differ in charge, topography, organic

functionality, and water structure.

The association of iodide and bromide with a complementary shaped hydrophobic

binding site suggests that the topography of protein surfaces (i.e., the shape of bumps, declivities,

or, perhaps, ion-binding motifs) may influence where, and how strongly, weakly hydrated ions

bind those surfaces, and highlights the inadequacy of continuum electrostatics models for

predicting ion-protein interactions. As with hydrophobic ligand-protein association, where

rearrangements of water and/or van der Waals interactions can contribute significantly to the

overall free energy of binding,26,48,74,76 anion association with the I-4 site is likely sensitive to the

  15  

local dielectric environment, which can differ significantly between (and within) binding

pockets.77,78 Accurate assessment of the prevalence and mechanistic basis of hydrophobic anion-

protein interactions will, thus, require additional crystallographic studies and thermodynamic

analyses of anion binding to different proteins.

Molecular dynamics simulations summarized in this work suggest an important

unanticipated effect of ions on the structure of water within concavities on protein surfaces.

Anions, upon associating with the Zn2+ cofactor, trigger rearrangements of water that extend well

beyond their first hydration shells (up to ~ 8 Å). This result suggests that concavities on surfaces

may amplify the distance over which ion-induced perturbations of water structure extend, to

distances well beyond that which characterizes the limit of their influence on the structure of

water in homogeneous solution. This amplification is consistent with the notion that water within

concavities on proteins exhibits long-range structure63—and, thus, long-range sensitivity to

perturbations—and suggests that ions bound to topographically complex surfaces may alter the

hydration state of residues beyond those immediately adjacent to their binding sites. We note,

however, that like the binding events themselves, ion-induced perturbations of water structure

are likely to be sensitive to local environment (e.g., electrostatics, protein topography) and, thus,

may differ significantly between binding pockets (and surfaces).

The results of this study suggest that the “Hofmeister series” describes what can be

considered to be—in the context of anion-protein association—a series of ligands. Even when

these ligands are identical in charge, they differ in their volume, shape, and affinity for water,

three attributes that strongly influence their ability to bind to—and alter the charge and hydration

structure of—polar, nonpolar, and topographically complex binding pockets.

  16  

Figure 1. The model system. (A) The Hofmeister series: anions ranked according to their

propensity to precipitate proteins from aqueous solution. In this study, we examined the

following anions: SO42-, HPO4

2-, CH3COO-, HCO3-, Cl-, Br-, NO3

-, I-, ClO4-, and SCN-. (B) The

association of anions with the Zn2+ cofactor involves two states: an initial state (left) with the

anion and protein in aqueous solution, and a final state (right) with the anion-protein complex in

aqueous solution. Thermodynamic parameters measured with ITC (ΔJ°bind, where J = H, TS, or

G) represent a difference between the initial and final states.

  17  

Figure 2. Thermodynamics of anion binding. (A) A plot showing thermodynamic parameters

for the association of anions and HCAII (298.15 K, pH = 7.6, 10 mM sodium phsophate buffer;

the process depicted in Figure 1B). H/S compensation, revealed by the plot, often arises from

rearrangements in the organization of waters that solvate interacting species. (B) A comparison

of free energies of hydration (ΔG°hydration) with free energies of binding (ΔG°bind,anion). Values of

ΔG°bind,anion decrease linearly with ΔG°hydration (R2 = 0.83), suggesting that anions with a lower

free energetic cost of dehydration bind more tightly to the Zn2+ cofactor. Values of ΔG°hydration

  18  

are taken from Marcus.36 Error bars represent standard error (n = 23 for the association of HCAII

and BTA in the absence of anions, and n ≥ 7 for the association of HCAII and BTA in the

presence of each anion; see SI Methods).

  19  

Figure 3. Structural basis of anion binding. (A) X-ray crystal structure of the active site of

HCAII complexed with SCN- (PDB entry 4YGK). Both ClO4- and SCN- displace H2O-338 (the

“so-called” deep water, displayed in Figure S2A) and shift the position of H2O-263 (the

catalytically important Zn2+-bound water). (B) X-ray crystal structure of the active site of HCAII

complexed with iodide (PDB entry 4YGN) sites are further elaborated in Appendix 3 of the SI).

Iodide sites are numbered in order of their proximity of the Zn2+-bound cofactor. I-1 and I-2

denote alternative binding sites for the Zn2+-bound iodide (an inner-sphere ion pair and an outer-

sphere ion pair, respectively). I-3 denotes a binding site at the border of the hydrophobic and

hydrophilic surfaces. I-4 denotes a binding site in the hydrophobic wall. Colors represent amino

acids as follows: cyan (within 5 Å of I-3), gray (within 5 Å of I-4), green (within 5 Å of both I-3

  20  

and I-4). (C) A detail of the I-3 binding site. Carbon atoms within 5 Å of the iodide are colored

cyan. (D) A detail of the I-4 binding site. Carbon atoms within 5 Å of the iodide are colored

gray. In both (C) and (D), the iodide atoms in the I-3 and I-4 positions, respectively, and the

Zn2+ cofactor are shown as spheres that indicate their solvent-accessible surface area (i.e., the

ion/water contact surface); the surface of the protein is also represented in this way.

  21  

  22  

Figure 4. Results from WaterMap calculations. (A) A plot showing the contribution of anion-

induced rearrangements of water inside the binding pocket of HCAII to the thermodynamics of

anion-Zn2+ association. Values of ΔJ°WM, anion represent the total difference of thermodynamic

properties (enthalpies, entropies, and free energies) of waters in anion-bound and anion-free

binding pockets (ΔJ°WM,anion = ΔJ°WM,HCA-anion - ΔJ°WM,HCA, where J = H, TS, or G). (B) A

schematic defining regions for calculating ΔH°WM,anion(d) and -TΔS°WM,anion(d), the enthalpy and

entropy, respectively, associated with rearrangements of water (resulting from anion-Zn2+

association) occurring beyond d Å from the surface of the Zn2+-bound anion (i.e., waters located

between a distance of d Å from the Zn2+-bound anion and the edge of the binding pocket). We

note: calculations are based on crystal structures of HCAII-anion complexes. (C) A plot showing

values of ΔH°WM-anion(d) and -TΔS°WM-anion(d) for the binding of SCN- to Zn2+. This plot suggests

that SCN- triggers rearrangements of water that extend up to 8 Å from its surface.

  23  

Figure 5. Rearrangements of water in the I-4 binding pocket. (A-B) WaterMap results for the

I-4 binding pocket shown in Figure 3D: (top) without iodide bound and (bottom) with iodide

bound. Waters are colored according to (A) their enthalpies (ΔH°WM) and (B) their entropies

(-TΔS°WM), relative to bulk water. In all images, the surface of the protein appears in gray.

Results suggest that the binding of iodide to the I-4 binding pocket causes displacement of two

enthalpically and entropically unstable (relative to bulk water) molecules of water (circled and

labeled with their corresponding thermodynamic quantities). In images, the surface of the protein

(gray) represents the protein/water contact surface.

  24  

ASSOCIATED CONTENT

Supporting Information. Experimental methods, appendices, and supporting figures, tables,

and references. This material is available free of charge at http://pubs.acs.org. Structure factors

and coordinates of anion-HCAII complexes are available in the Protein Data Bank

(www.rcsb.org) with reference codes 4YGK, 4YGL, 4YGN, and 4YGJ.

AUTHOR INFORMATION

Corresponding Author. *gwhitesides@gmwgroup.harvard.edu

Present Addresses. † Department of Chemistry, University of North Carolina at Chapel Hill,

Chapel Hill, NC 27599.

Notes. The authors declare no competing financial interests.

ACKNOWLEDGEMENTS

We would like to thank Serena Bai and Phil Snyder (Harvard University) for initial X-ray

crystallography studies of the association of iodide and the binding pocket of HCAII, and

Benjamin Breiten (Harvard University) for helpful discussions. This work was supported by the

National Science Foundation under Award No. 1152196. A.H was additionally supported by

NIH/NIGMS grant 8P41GM103473-16 and DOE/BER grant BO-70. Structural data was

collected at beamline X25 of the National Synchrotron Light Source under support of DOE/BES

contract No. DE-AC02-98CH10886.

  25  

REFERENCES

(1) Eveloff, J. L.; Warnock, D. G. Am. J. Physiol. 1987, 252, F1.

(2) Sterling, D.; Casey, J. R. Biochem. Cell Biol. 2002, 80, 483.

(3) Desai, S. A.; Bezrukov, S. M.; Zimmerberg, J. Nature 2000, 406, 1001.

(4) Hedrich, R. Neher, E. Nature 1987, 329, 833.

(5) Takahashi, H.; Yamazaki, M.; Sasakura, N.; Watanabe, A.; Leustek, T.; Engler, J. A.; Engler, G.; Van Montagu, M.; Saito, K. Proc. Natl. Acad. Sci. U. S. A. 1997, 94, 11102.

(6) Clapham, D. E. Cell 2007, 131, 1047.

(7) Jasti, J.; Furukawa, H.; Gonzales, E. B.; Gouaux, E. Nature 2007, 449, 316.

(8) Catterall, W. A. Science 1988, 242, 50.

(9) Sanguinetti, M. C.; Tristani-Firouzi, M. Nature 2006, 440, 463.

(10) McLachlan, S. M.; Rapoport, B. Endocr. Rev. 2013, 13, 192.

(11) Hunter, T. Cell 1995, 80, 225.

(12) Tobias, D. J.; Hemminger, J. C. Science (80-. ). 2008, 319, 1197.

(13) Jungwirth, P.; Cremer, P. S. Nat. Chem. 2014, 6, 261.

(14) Vrbka, L.; Vondrásek, J.; Jagoda-Cwiklik, B.; Vácha, R.; Jungwirth, P. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15440.

(15) Collins, K. D. Biophys. Chem. 2012, 167, 33.

(16) Nostro, P. Lo; Ninham, B. Chem. Rev. 2012, 112, 2286.

(17) Moreira, L. A.; Boström, M.; Ninham, B. W.; Biscaia, E. C.; Tavares, F. W. Colloids Surfaces A Physicochem. Eng. Asp. 2006, 282-283, 457.

(18) Collins, K. D. Biophys. Chem. 2006, 119, 271.

(19) Heyda, J.; Hrobárik, T.; Jungwirth, P. J. Phys. Chem. A 2009, 113, 1969.

(20) Vrbka, L.; Jungwirth, P.; Bauduin, P.; Touraud, D.; Kunz, W. J. Phys. Chem. B 2006, 110, 7036.

  26  

(21) Zhang, Y.; Furyk, S.; Bergbreiter, D. E.; Cremer, P. S. J. Am. Chem. Soc. 2005, 127, 14505.

(22) Chen, X.; Yang, T.; Kataoka, S.; Cremer, P. S. J. Am. Chem. Soc. 2007, 129, 12272.

(23) Kalra, A.; Tugcu, N.; Cramer, S. M.; Garde, S. J. Phys. Chem. B 2001, 105, 6380.

(24) Zhang, Y.; Cremer, P. S. 2006, 10, 658.

(25) Krishnamurthy, V. M.; Kaufman, G. K.; Urbach, A. R.; Gitlin, I.; Gudiksen, K. L.; Weibel, D. B.; Whitesides, G. M. Chem. Rev. 2008, 108, 946.

(26) Snyder, P. W.; Lockett, M. R.; Moustakas, D. T.; Whitesides, G. M. Eur. Phys. J. Spec. Top. 2014, 223, 853.

(27) Eriksson, A. E.; Kylsten, P. M.; Jones, T. A.; Liljas, A. Proteins 1988, 4, 283.

(28) Christianson, D. W.; Fierke, C. A. Acc. Chem. Res. 1996, 29, 331.

(29) Mazumdar, P. A.; Kumaran, D.; Swaminathan, S.; Das, A. K. Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. 2008, 64, 163.

(30) Jonsson, B. M.; Hakansson, K.; Liljas, A. Febs Lett. 1993, 322, 186.

(31) Mangani, S.; Hakansson, K. Eur. J. Biochem. 1992, 210, 867.

(32) Kannan, K. K.; Ramanadham, M.; Jones, T. A. Ann. N. Y. Acad. Sci. 1984, 429, 49.

(33) Fisher, Z.; Boone, C. D.; Biswas, S. M.; Venkatakrishnan, B.; Aggarwal, M.; Tu, C.; Agbandje-Mckenna, M.; Silverman, D.; McKenna, R. Protein Eng. Des. Sel. 2012, 25, 347.

(34) Zhang, Y.; Cremer, P. S. Annu. Rev. Phys. Chem. 2010, 61, 63.

(35) Collins, K. D. Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 5553.

(36) Marcus, Y. Ion Properties; Marcel Dekker: New York, New York, 1997.

(37) Jungwirth, P.; Winter, B. Annu. Rev. Phys. Chem. 2008, 59, 343.

(38) Collins, K. D. Biophys. J. 1997, 72, 65.

(39) Hess, B.; van der Vegt, N. F. A. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 13296.

(40) Collins, K. D. Methods 2004, 34, 300.

  27  

(41) Uejio, J. S.; Schwartz, C. P.; Duffin, A. M.; Drisdell, W. S.; Cohen, R. C.; Saykally, R. J. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 6809.

(42) Ries-Kautt, M.; Ducruix, A. Methods Enzymol. 1997, 276, 23.

(43) Kherb, J.; Flores, S. C.; Cremer, P. S. J. Phys. Chem. B 2012, 116, 7389.

(44) Collins, K. D.; Neilson, G. W.; Enderby, J. E. Biophys. Chem. 2007, 128, 95.

(45) Zhang, Y. L.; Zhang, Z. Y. Anal. Biochem. 1998, 261, 139.

(46) Snyder, P. W.; Mecinovic, J.; Moustakas, D. T.; Thomas, S. W.; Harder, M.; Mack, E. T.; Lockett, M. R.; Heroux, A.; Sherman, W.; Whitesides, G. M. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 17889.

(47) Dunitz, J. D. Chem. Biol. 1995, 2, 709.

(48) Breiten, B.; Lockett, M. R.; Sherman, W.; Fujita, S.; Al-Sayah, M.; Lange, H.; Bowers, C. M.; Heroux, A.; Krilov, G.; Whitesides, G. M. J. Am. Chem. Soc. 2013, 135, 15579.

(49) Chodera, J. D.; Mobley, D. L. Annu. Rev. Biophys. 2013, 42, 121.

(50) Sjöblom, B.; Polentarutti, M.; Djinovic-Carugo, K. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 10609.

(51) Lund, M.; Vácha, R.; Jongwirth, P. Langmuir 2008, 24, 3387.

(52) Lund, M.; Vrbka, L.; Jungwirth, P. J. Am. Chem. Soc. 2008, 130, 11582.

(53) Huang, D. M.; Cottin-Bizonne, C.; Ybert, C.; Bocquet, L. Phys. Rev. Lett. 2007, 98, 177801.

(54) Horinek, D.; Netz, R. R. Phys. Rev. Lett. 2007, 99, 226104.

(55) Gibb, C. L. D.; Gibb, B. C. J. Am. Chem. Soc. 2011, 133, 7344.

(56) Carnegie, R. S.; Gibb, C. L. D.; Gibb, B. C. Angew. Chemie Int. Ed. 2014, 53, 11498.

(57) Pegram, L. M.; Wendorff, T.; Erdmann, R.; Shkel, I.; Bellissimo, D.; Felitsky, D. J.; Record, M. T. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 7716.

(58) Dauter, Z.; Dauter, M.; de La Fortelle, E.; Bricogne, G.; Sheldrick, G. M. J. Mol. Biol. 1999, 289, 83.

(59) Dauter, Z.; Dauter, M.; Rajashankar, K. R. Acta Crystallogr. Sect. D Biol. Crystallogr. 2000, 56, 232.

  28  

(60) Baker, N. A.; Sept, D.; Joseph, S.; Holst, M. J.; McCammon, J. A. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 10037.

(61) Pegram, L. M.; Record, M. T. J. Phys. Chem. B 2008, 112, 9428.

(62) Pegram, L. M.; Record, M. T. J. Phys. Chem. B 2007, 111, 5411.

(63) Young, T.; Abel, R.; Kim, B.; Berne, B. J.; Friesner, R. A. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 808.

(64) Abel, R.; Young, T.; Farid, R.; Berne, B. J.; Friesner, R. A. J. Am. Chem. Soc. 2008, 130, 2817.

(65) Beuming, T.; Farid, R.; Sherman, W. Protein Sci. 2009, 18, 1609.

(66) Lazaridis, T. J. Phys. Chem. B 1998, 102, 3531.

(67) Lazaridis, T. J. Phys. Chem. B 1998, 102, 3542.

(68) Omta, A. W.; Kropman, M. F.; Woutersen, S.; Bakker, H. J. Science 2003, 301, 347.

(69) Wachter, W.; Kunz, W.; Buchner, R.; Hefter, G. J. Phys. Chem. A 2005, 109, 8675.

(70) Buchner, R.; Hefter, G. T.; May, P. M. J. Phys. Chem. A 1999, 103, 1.

(71) Boström, M.; Williams, D. R. M.; Ninham, B. W. Langmuir 2001, 17, 4475.

(72) Otten, D. E.; Shaffer, P. R.; Geissler, P. L.; Saykally, R. J. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 3190.

(73) Caleman, C.; Hub, J. S.; van Maaren, P. J.; van der Spoel, D. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 6838.

(74) Barratt, E.; Bingham, R. J.; Warner, D. J.; Laughton, C. A.; Phillips, S. E. V; Homans, S. W. J. Am. Chem. Soc. 2005, 127, 11827.

(75) Homans, S. W. Drug Discov. Today 2007, 12, 534.

(76) Bingham, R. J.; Findlay, J. B. C.; Hsieh, S.-Y.; Kalverda, A. P.; Kjellberg, A.; Perazzolo, C.; Phillips, S. E. V; Seshadri, K.; Trinh, C. H.; Turnbull, W. B.; Bodenhausen, G.; Homans, S. W. J. Am. Chem. Soc. 2004, 126, 1675.

(77) Li, L.; Li, C.; Zhang, Z.; Alexov, E. J. Chem. Theory Comput. 2013, 9, 2126.

(78) Dwyer, J. J.; Gittis, A. G.; Karp, D. A.; Lattman, E. E.; Spencer, D. S.; Stites, W. E.; García-Moreno E, B. Biophys. J. 2000, 79, 1610.

  29  

TABLE OF CONTENTS